Automatic Ischemic Stroke Lesions Segmentation in Multimodality MRI using Mask Region-based Convolutional Neural Network

Stroke or cerebrovascular accident (CVA) disease is one of the leading causes of death, due to its difficult diagnosis. The speed of its treatment has a direct impact on patients' lives. Acute ischemic lesions occur in most CVA patients. Although FLAIR and diffusion-weighted MR imaging (DWI) ar...

Full description

Saved in:
Bibliographic Details
Published in2020 4th International Conference on Advanced Systems and Emergent Technologies (IC_ASET) pp. 362 - 366
Main Authors Daoudi, Rimeh, Mouelhi, Aymen, Sayadi, Mounir
Format Conference Proceeding
LanguageEnglish
Published IEEE 15.12.2020
Subjects
Actue ischemic stroke
Convolutional neural networks
deep learning
Diseases
Image segmentation
Lesions
Magnetic resonance imaging
medical image segmentation
Predictive models
Solid modeling
Training
Online AccessGet full text

Cover

More Information
Summary:Stroke or cerebrovascular accident (CVA) disease is one of the leading causes of death, due to its difficult diagnosis. The speed of its treatment has a direct impact on patients' lives. Acute ischemic lesions occur in most CVA patients. Although FLAIR and diffusion-weighted MR imaging (DWI) are sensitive to these lesions, localizing and assessing them manually is time consuming and challenging for clinicians. In this paper, we present an effective method to detect and segment stroke lesions in multimodal MR images using mask region-based convolutional neural network (MASK R-CNN). It is validated on a large dataset comprising clinical acquired multimodal MR images including FLAIR, T2 and DWI from 37 subjects. The mean average precision (mAP) metric based on testing subjects with small and large lesions is 0.81.
DOI:10.1109/IC_ASET49463.2020.9318265